Enantiospecific Synthesis of Inositol Polyphosphates

May 30, 1991 - An efficient and versatile strategy for the preparation of highly differentiated cyclitols from (-)-quinic acid was developed and explo...
0 downloads 0 Views 835KB Size
Chapter 11

Downloaded by UNIV MASSACHUSETTS AMHERST on August 2, 2012 | http://pubs.acs.org Publication Date: May 30, 1991 | doi: 10.1021/bk-1991-0463.ch011

Enantiospecific Synthesis of Inositol Polyphosphates, Phosphatidylinositides, and Analogues from (—)-Quinic Acid J. R. Falck and Abdelkrim Abdali Department of Molecular Genetics and Pharmacology, University of Texas Southwestern Medical Center, Dallas, TX 75235

An efficient and versatile strategy for the preparation of highly differentiated cyclitols from (-)-quinic acid was developed and exploited for the enantiospecific total synthesis of representative inositol polyphosphates and phosphatidylinositides. The 5methylenephosphonate analogue of D-myo-inositol 1,4,5trisphosphate was also prepared and shown to be a long-lived agonist of calcium mobilization. In recognition of the physiologic significance of phosphorylated cyclitols and their limited availability from natural sources, there has been a resurgence of interest in the chemical synthesis of functionalized inositols and analogues (1,2). Until recently, virtually all preparative studies utilized myo-inositol as the initial precursor and relied upon sometimes circuitous protection/deprotection sequences to differentiate amongst this cyclitoPs six hydroxyls. Furthermore, since the myo-isomer is a meso form of inositol, a resolution is required to obtain optically active products. These limitations have been partially addressed (3-5) using pinitol and quebrachitol, two chiral, but not widely available natural cyclitols. Efforts to exploit carbohydrates have also been reported (6). It should be noted that the novel intermediate, cw-l,2-dihydroxycyclohexa-3,5-diene, arising from microbial oxidation of benzene, has also been used as the starting point for inositol syntheses (7,8). Given the increasing popularity of enzymatic methodology in organic synthesis, it is reasonable to anticipate additional contributions utilizing such procedures. As part of a comprehensive program in these laboratories for the stereorational total synthesis of inositol polyphosphates and associated phospholipids, we developed a conceptually different, and potentially more flexible, approach to functionalized cyclitols that exploits a relatively inexpensive and readily available member of the chiral pool, (-)-quinic acid (1). Our strategy also offers, in principle, greater latitude for preparing configurational and constitutional analogues.

0097-6156/91/0463-0145$06.00/0 © 1 9 9 1 American Chemical Society

In Inositol Phosphates and Derivatives; Reitz, A.; ACS Symposium Series; American Chemical Society: Washington, DC, 1991.

146

INOSITOL

PHOSPHATES

AND DERIVATIVES

Downloaded by UNIV MASSACHUSETTS AMHERST on August 2, 2012 | http://pubs.acs.org Publication Date: May 30, 1991 | doi: 10.1021/bk-1991-0463.ch011

Inositol 1,4,5-Trisphosphate Our initial synthetic target was D-myo-inositol 1,4,5-trisphosphate (10), the calcium-mobilizing intracellular second messenger of the phosphatidylinositol (PI) cycle (9). Of the possible perspectives from which to view 1 with respect to the myo-inositol moiety 2 present in 10, the orientation depicted in la provides the greatest topologic congruity, i.e., the three contiguous quinate alcohols have the same relative and absolute configurations as C(2), C(l), and C(6), respectively, when la is superimposed upon 2. Projection of orientation lb reveals coincidence with only two stereocenters, those at C(2) and C(3). The third alcohol at C(l) is epimeric. Yet, despite its lesser stereochemical homology, other considerations (vide infra) made lb the more attractive candidate provided satisfactory solutions could be found for several outstanding issues. These included: (i) differentiation of the hydroxyls, (ii) stereospecific oxidation of both methylenes corresponding to C(4) and C(6), and (iii) onecarbon degradation to remove the carboxyl. The realization (10) of these objectives culminating in an efficient total synthesis of 10 is outlined in Scheme I. Firstly, the C(l)-stereochemistry was adjusted using a modification of literature procedure that improved the overall yield and simplified isolation of intermediates. Specifically, (-)-quinic acid (1) was converted to lactone 3 by concurrent lactonization/ ketalization utilizing cyclohexanone and Amberlite IR120 resin followed by mesylation of the tertiary alcohol under standard conditions. Sequential lactone methanolysis, pyridinium chlorochromate (PCC) oxidation, and Et N induced mesylate elimination generated an enone from which ester 4 was obtained by hydride delivery exclusively from the less hindered β-face. Taking advantage of the residual functionality created above, the C(6)hydroxyl was introduced stereospecifically with the desired configuration. For this, ester 4 was transformed to phenyl selenide 5 by protection of the C(l)alcohol as its B-trimethylsilylethoxymethyl(SEM) ether, reduction of the ester using diisobutylaluminum hydride (DIBAL-H) at -78°C, and selenylation of the resultant primary alcohol with N-(phenylseleno)phthalimide in the presence of Bu P (11). Stereoselective in situ [2,3]-sigmatropic rearrangement of the allylic selenoxide derived from 5 and benzylation of the product evolved 6 as the sole product. Due to the high cost of the selenating reagent, the sulfide version of 5 was also examined. In this instance, peracid oxidation yielded an ca. 1:1 mixture of sulfoxides that led in refluxing benzene to a mixture of epimeric alcohols ( α / β 1:3); at 45°C, however, only the β-isomer was observed. The favorable outcome of this rearrangement can be attributed to (i) facile racemization of the sulfoxide via rapid equilibration with its sulfinate ester and (ii) preferential approach of the thiophile from the β-face. Inspection of molecular models confirms the α-side is well-screened by the bulky cyclohexylidene substituent (cf.20). To functionalize the remaining methylene, 6 was subjected to Se0 allylic oxidation (78%), then ozonolysis (83%) in M e O H / C H C l . However, hydride 3

3

2

2

2

In Inositol Phosphates and Derivatives; Reitz, A.; ACS Symposium Series; American Chemical Society: Washington, DC, 1991.

11.

F A L C K & ABDALI

Enantiospecific Synthesis from (-)-Quinic Acid

147

Scheme I

1. NaOMe, MeOH 2. PCC; Et N 3

Downloaded by UNIV MASSACHUSETTS AMHERST on August 2, 2012 | http://pubs.acs.org Publication Date: May 30, 1991 | doi: 10.1021/bk-1991-0463.ch011

(-)-Quinic Acid

3. N a B H , 0 ° C 80%

2. M s C l , E t N , 0 ° C 85%

4

3

1

SePh C0 Me 2

l . S E M - α , (iPr^NEt 2. DIBAL-H, PhCH -78°C

1. NaI0 ,pH 7 buffer, 4

3

1,4-Dioxane, 0°C 2. K H , BnBr

3. N-(PhSe)phth, Bu P, -15°C 3

42%

1. BH ,THF, 0°C;

1.0 , C H a / M e O H 3

2

3

2

Alk. H Q

-78°C

2

4

3

0

91%

OSEM

OSEM

(HO^OPO,

1. K H , THF, 60° C [(BnO) PO] Q 2

2

2. H ,50psig, 10%Pd/C, 95% EtOH ; AcOH/H 0 2

2

\>PO(OH)

62%

2

8 :R =TBDMS R»=SEM 9 :R=R'=H

2

2. n-Bu NF, HMPA, 4 A M.S., 100°C 72 %

2. TBDMSOTf, E l N

M,5-IP j 3

American Chemical Society Library 1155 16th St. N.W. f

In Inositol Phosphates andD.C. Derivatives; Reitz, A.; Washington, 20036 ACS Symposium Series; American Chemical Society: Washington, DC, 1991.

148

INOSITOL

PHOSPHATES

A N D DERIVATIVES

reduction of the ketone 11, thus obtained, afforded a separable mixture of C(5)diastereomeric alcohols under a variety of conditions. Efforts to epimerize the undesired isomer proved disappointing. In contrast, ozonolysis of 6 and conversion to silyl enol ether 7 using excesstert-butyldimethylsilyl(TBDMS) triflate according to Corey (12) proceeded smoothly and with virtually complete regiospecificity. Hydroboration of the enol ether from the more accessible β-face followed by careful alkaline peroxide oxidation furnished differentially protected cyclitol 8. While the TBDMS ether was easily severed, more drastic conditions (n-Bu NF,HMPA,100°C) were needed to coerce the SEM in order to obtain triol 9. The identity of 9 was confirmed by comparisons of its physical [mmp 136137°C; lit. (13) 137-139°C] and spectral characteristics with an authentic sample. Final elaboration of 9 with tetrabenzyl pyrophosphate and removal of the protecting groups as described by Vacca (14) provided D-myo-inositol 1,4,5trisphosphate 10, isolated as its hexasodium salt.

Downloaded by UNIV MASSACHUSETTS AMHERST on August 2, 2012 | http://pubs.acs.org Publication Date: May 30, 1991 | doi: 10.1021/bk-1991-0463.ch011

4

5-Methylenephosphonate Analogue of 1,4,5-IP

3

In vivo, 1,4,5-IP (10) is rapidly metabolized by either of two divergent pathways: (a) initial phosphorylation by a 3-kinase and subsequent cleavage of the C(5)phosphate, or (b) direct dephosphorylation by a specific 5-phosphatase (1,2). Intervention in these pathways may provide new insights into the function and possible therapeutic manipulation of the PI cycle. Special emphasis has been placed on the development of modified inositol polyphosphates with metabolically more stable phosphorus functionalities (15-19). Consequently, it was of interest to prepare 14, the 5-methylenephosphonate analogue (20) of 1,4,5-IP , by an extension of our strategy that utilizes both the chirality and complete carbon framework of (-)-quinic acid (Scheme II). Fluoride ion induced desilylation of 6 in hexamethylphosphoric triamide (HMPA) and kinetically (21) controlled addition of phenylselenenyl bromide across the exocyclic alkene provided mainly the anti-Markovnikov adduct. The latter underwent regiospecific oxidative elimination to the somewhat labile allylic bromide 12 that was sufficiently pure for Michaelis-Becker phosphorylation (22) using sodium dibenzyl phosphite (generated in situ from NaH and dibenzyl phosphite) in the presence of 18-crown-6 to improve the solubility of the reagent in toluene. Subsequent hydroboration of the alkene followed by oxidative work­ up using peracid, rather than alkaline H 0 which hydrolyzes the phosphonate, furnished diol 13 with the desired M-trans configuration between the substituents at C(3)-C(6) (*Η NMR analysis: COSY, J-resolved). Introduction of the C(l) and C(4) phosphates by the two-step phosphite method (3) of Tegge and Ballou and removal of the protecting groups afforded 14, isolated as its sodium salt. Phosphonate 14 elicits contraction of bovine tracheal smooth muscle (25), permeabilized with saponin, with an efficacy five to ten times less than a comparable concentration of 1,4,5-IP . Evaluation (24) of its ability to displace [ P]-1,4,5-IP bound to bovine adrenocortical membranes showed 14 has an affinity for the 1,4,5-IP binding site ca. two orders of magnitude less than unlabeled 1,4,5-IP (Figure 1). For release of C a from bovine adrenal 3

3

2

2

3

32

3

3

2+

3

In Inositol Phosphates and Derivatives; Reitz, A.; ACS Symposium Series; American Chemical Society: Washington, DC, 1991.

Downloaded by UNIV MASSACHUSETTS AMHERST on August 2, 2012 | http://pubs.acs.org Publication Date: May 30, 1991 | doi: 10.1021/bk-1991-0463.ch011

11. FALCK & ABDALI

Enantiospecific Synthesisfrom(-)-Quinic Acid

149

1. NaP0(0Bn)2 ;18-Crown-6

6

2. BH ,THF ; mCPBA 62% 3

PO(OBn)

"

2

OBn

1. (iPr) NP(OBn) , 1-H-Tetrazole; mCPBA ^ 2. H ,50 psig, 10% Pd/C, 80% EtOH; AcOH/H 0 61% 2

(HO)jOPO

2

2

2

5-Methylenephosphonate Analogue

In Inositol Phosphates and Derivatives; Reitz, A.; ACS Symposium Series; American Chemical Society: Washington, DC, 1991.

150

INOSITOL

PHOSPHATES

A N D DERIVATIVES

microsomal preparations (24), as measured by Fura 2 fluorescence, 14 and 2,4,5IP were nearly equally potent and about one-fifth as active as the natural second messenger, 1,4,5-IP (Figure 2). In contrast with the latter compound, 14 and 2,4,5-IP give an initial sharp increase in the unbound C a concentration that then remains elevated above basal values (24). These data are consistent with an attenuated, but long-lived, 1,4,5-IP agonist. 3

3

2+

3

3

Downloaded by UNIV MASSACHUSETTS AMHERST on August 2, 2012 | http://pubs.acs.org Publication Date: May 30, 1991 | doi: 10.1021/bk-1991-0463.ch011

A Second Generation Approach To date, almost two dozen inositol phosphate and phosphatidylinositide isomers have been isolated and identified from natural sources (7,2). Although recent investigations have established a complex interrelationship amongst these metabolites, the physiological role(s) for all but a few remains obscure. The more recent discovery (25,26) of a metabolically distinct class of 3phosphorylated phosphatidylinositides portends even greater diversity. To help encompass this growing synthetic challenge, we sought to extend the scope of our basic theme by preparing a more highly differentiated intermediate and then illustrating its usefulness in the synthesis of some representative examples of current interest. A convenient point of departure was lactone 15 available (27) on a large scale in 50-60% yield from 1 by vacuum sublimation (equation 1). The corresponding cyclic dibutyl stannylene ether, generated (28) in situ from 15, was enticed by an equivalent amount of n-Bu NI to react with SEM-C1 at 60°C to give diol 16 (78%). None of the regioisomeric SEM ether could be detected by chromatographic analysis. Yet, introduction of a benzyl or silyl protecting group at the adjacent alcohol proved unexpectedly troublesome. Unacceptable product mixtures were obtained due to the high reactivity of the tertiary alcohol. This approach was ultimately abandoned for one that called into service enone 17, an intermediate in the preceding syntheses. Concurrent reduction of the ester and keto functionalities in 17 with DIBAL-H followed by selective replacement of the primary alcohol using phenyl disulfide/tributylphosphine was rewarded by a good yield of phenylsulfide 18 (Scheme III). A 3,4-dimethoxybenzyl was staked onto the remaining secondary alcohol and the cyclohexylidene was removed under mild acidic conditions. Tin mediated etherification of the resultant diol 19 with SEM-C1 using nBu NI (vide supra) showed lower regioselectivity (4:1) for the equatorial C(3)-alcohol than did CsF catalysis (24:1). Subsequent benzylation leading to 20 was uneventful. Low temperature peracid oxidation gave a - 1:1 mixture of sulfoxides that was transformed into olefin 21, free of any α-isomer, by [2,3]-sigmatropic rearrangement and benzylation of the newly generated allylic alcohol. In this instance, rearrangement required a significantly longer reaction time and bulkier thiophile than the comparable cyclohexylidene examples described above. Having observed that the 3,4-dimethoxybenzyl ether did not survive ozonolytic cleavage of the exocyclic olefin, ketone 22 was secured by an alternative twostep, but very efficient, sequence involving Os0 glycolization and Pb(OAc) cleavage. Proceeding with 22 along the now well established route of enol ether 4

4

4

In Inositol Phosphates and Derivatives; Reitz, A.; ACS Symposium Series; American Chemical Society: Washington, DC, 1991.

4

11.

F A L C K & ABDALI

Enantiospecific Synthesisfrom(-)-Quinic Acid

Downloaded by UNIV MASSACHUSETTS AMHERST on August 2, 2012 | http://pubs.acs.org Publication Date: May 30, 1991 | doi: 10.1021/bk-1991-0463.ch011

120

.01

.1

1

10

100

1000

[INOSITOL PHOSPHATE] 32

Figure l. P-Ins(l,4,5)P displacement in bovine adrenal microsomal membranes. 3

Figure 2. Fura 2 measurement of calcium release in bovine adrenal microsomes.

In Inositol Phosphates and Derivatives; Reitz, A.; ACS Symposium Series; American Chemical Society: Washington, DC, 1991.

151

152

Downloaded by UNIV MASSACHUSETTS AMHERST on August 2, 2012 | http://pubs.acs.org Publication Date: May 30, 1991 | doi: 10.1021/bk-1991-0463.ch011

INOSITOL PHOSPHATES AND DERIVATIVES

Scheme III COjMe l.NaH.RBr

1. DIBAL-H, PhCH , -78°C 3

2. H Q , MeOH

2. PhSSPh,Bu P,0°C 3

80%

72%

1. nBu SnO, PhCH , 120°C; SEMC1, CsF, -IS to 0°C 2

3

1. mCPBA; (EtO^P, MeOH 2. K H , BnBr

2. KH, BnBr

83%

66%

19

1.0s0 ,NMO

1. TBDMSOTf, Et N

2. Pb(OAc)

2. Bfy Me^S ; mCPBA 66%

3

4

4

90%

22

21

O i TB DMS HQ.

!

OPO(OBn)2 (BnOjOPO^

JDBn

.OBn

l.HCl.MeOH 2. (iPr) NP(OBn) ,1-H-Teirazole; mCPBA _^ ^ 3. DDQ, CH Cl /pH7 buffer (BnO^oi 2

2

m

2

x

2

OBn 24

23 R= 3,4-(MeO) C6H CH 2

3

2

In Inositol Phosphates and Derivatives; Reitz, A.; ACS Symposium Series; American Chemical Society: Washington, DC, 1991.

Enantiospecific Synthesisfrom(-)-Quinic Acid

FALCK & ABDALI

11.

153

Scheme IV

24

1. (iPr^NPiOBiOj, 1-H-Tetrazole; mCPBA Ζ H , 50 psig, 10% Pd/C, / 80% EtOH 2

ι—OR 90%

j

OBn 2. H . 50 psig, 10% Pd/C, 80% EtOH

H ,50 psig 10% Pd/C, 80% EtOH

Downloaded by UNIV MASSACHUSETTS AMHERST on August 2, 2012 | http://pubs.acs.org Publication Date: May 30, 1991 | doi: 10.1021/bk-1991-0463.ch011

2

OPO(OH), (HO),OPO